摘要
为解决航空集群网络(ASNET)利用认知抗干扰频谱接入时会发生信道碰撞从而降低通信性能问题,研究了基于多臂赌博机(MAB)理论的航空认知抗干扰频域信道选择技术.首先,构建航空集群网络抗干扰信道选择MAB博弈模型,给出了准确估算动态集群网络电台数量的算法;然后,基于此先验信息提出碰撞规避(CA)的klUCB++抗干扰信道选择策略,并进一步推导出信道碰撞次数的理论上界.仿真结果表明:所提出的CA kl-UCB++抗干扰信道选择策略降低了电台频谱接入的碰撞概率和累积悔值(regret),能够有效提高航空集群网络的频域抗干扰通信性能.
To solve the problem that multiple airborne radios access to the same channel would result in collision and degrade the system performance in aeronautical swarm network(ASNET),the cognitive anti-jamming channel selection scheme based on multiarmed bandit(MAB) learning was investigated. The ASNET anti-jamming channel selection MAB game model was constructed firstly, and the accurate algorithm was given to estimate the number of radios in dynamic swarm networks. Then, a collision avoidance(CA) kl-UCB++anti-jamming channel selection strategy was proposed based on the prior estimated information.Furthermore, the theoretical upper bound of the number of channel collisions was derived. Simulation results show that the proposed CA kl-UCB++anti-jamming channel selection strategy can reduce the collisions on available channels with lower cumulative regret,which can effectively improve the ASNET anti-jamming ability in frequency domain.
作者
仇启明
黎海涛
张昊
罗佳伟
QIU Qiming;LI Haitao;ZHANG Hao;LUO Jiawei(China National Aeronautical Radio Electronics Institute.Shanghai 200233,China;Faculty of Information Technology,Beijing University of Technology,Beijing 100124,China)
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2021年第5期20-25,共6页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
航空科学基金资助项目(2018ZC15003)。
